/
SHT.py
227 lines (180 loc) · 7.86 KB
/
SHT.py
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
import sys
from pathlib import Path
path = Path(__file__).parent
sys.path.append(str(path.resolve()) + '/obj/')
import fastSHT
import numpy as np
import healpy as hp
import time
import numba
from numba import cuda
class SHT:
def __init__(self, nside, lmax, nsim, niter=0, pol=False, all_buff=False):
self.nside = nside
self.lmax = lmax
self.nsim = nsim
self.niter = niter
if(pol == True):
self.pol = 1
else:
self.pol = 0
self.nring = 4 * nside - 1
self.npix = 12 * nside ** 2
ring_tab_dir = str(path.resolve()) + '/SHT_ring_skip_table/'
ring_tab = np.fromfile(ring_tab_dir + str(nside).rjust(4,'0') + '.bin', dtype=np.int32)
plm_dir = str(path.resolve()) + '/plm_table/'
plm_pos = np.fromfile(plm_dir + 'plm_pos_' + str(nside).rjust(4,'0') + '.bin', dtype=np.int32)
plm_val1 = np.fromfile(plm_dir + 'plm_val1_' + str(nside).rjust(4,'0') + '.bin', dtype=np.double)
plm_val2 = np.fromfile(plm_dir + 'plm_val2_' + str(nside).rjust(4,'0') + '.bin', dtype=np.double)
plm_pos = np.asfortranarray(plm_pos.reshape(2*nside, int(len(plm_pos) / (2*nside)), order='F')[:,:lmax+1], dtype=np.int32)
plm_val1 = np.asfortranarray(plm_val1.reshape(2*nside, int(len(plm_val1) / (2*nside)), order='F')[:,:lmax+1], dtype=np.double)
plm_val2 = np.asfortranarray(plm_val2.reshape(2*nside, int(len(plm_val2) / (2*nside)), order='F')[:,:lmax+1], dtype=np.double)
plm_val1[plm_pos >= lmax] = 0
plm_val2[plm_pos >= lmax] = 0
plm_pos[plm_pos >= lmax] = lmax-1
tmp = np.concatenate([np.arange(4,4*nside,4), [4 * nside]*(2*nside-1), np.flip(np.arange(4,4*nside,4))])
TJI = np.concatenate([[0], np.cumsum(tmp)])
theta,phi0 = hp.pix2ang(nside, TJI)
self.nbuff = 1
if pol == True or niter > 0:
self.nbuff = 2
if pol == True and niter > 0:
self.nbuff = 4
if all_buff == True:
self.nbuff = 4
fastSHT.sht_data_alloc(np.array((self.nside, self.lmax, self.nring, self.nsim, self.nbuff, self.pol), dtype=np.int32))
fastSHT.sht_set_data(ring_tab[:lmax+1],plm_pos, plm_val1, plm_val2, theta, phi0)
# set the map in the SHT class
def set_maps(self, maps_in):
self.maps = maps_in
# set the alms in the SHT class
def set_alms(alms_in):
self.alms = alms_in
def t2alm(self, maps_in=None, alms_in=None, method='old', get_alm=True):
if(maps_in is None):
maps = self.maps
else:
maps = maps_in
if(maps is None):
raise ValueError('No input maps AND no internal maps! Check the parameters or call set_maps(maps_in)!')
if(alms_in is None and get_alm == True):
try:
alms = numba.cuda.pinned_array((self.nsim, self.lmax+1, self.lmax+1), dtype=np.double, strides=None, order='F')
except:
alms = np.ones((self.nsim, self.lmax+1, self.lmax+1), dtype=np.double, order='F')
else:
alms = alms_in
if(get_alm == False):
if(self.niter == 0):
fastSHT.t2alm(maps)
elif(method == 'old'):
fastSHT.t2alm_iter_old(maps, self.niter)
else:
fastSHT.t2alm_iter(maps, self.niter)
return
else:
if(self.niter == 0):
fastSHT.t2alm(maps, alms)
elif(method == 'old'):
fastSHT.t2alm_iter_old(maps, self.niter, alms)
else:
fastSHT.t2alm_iter(maps, self.niter, alms)
if(alms_in is None):
return alms
else:
return 0
def alm2t(self, alms, maps_in=None):
if(maps_in is None):
try:
maps = numba.cuda.pinned_array((self.npix, self.nsim), dtype=np.double, strides=None, order='F')
except:
maps = np.ones((self.npix, self.nsim), dtype=np.double, order='F')
else:
maps = maps_in
fastSHT.alm2t( alms, maps)
if(maps_in is None):
return maps
else:
return 0
def qu2eb(self, Q, U, almEs_in=None, almBs_in=None, method='old'):
if(almEs_in is None or almBs_in is None):
try:
almEs = numba.cuda.pinned_array((self.nsim, self.lmax+1, self.lmax+1), dtype=np.double, strides=None, order='F')
almBs = numba.cuda.pinned_array((self.nsim, self.lmax+1, self.lmax+1), dtype=np.double, strides=None, order='F')
except:
almEs = np.ones((self.nsim, self.lmax+1, self.lmax+1), order='F')
almBs = np.ones((self.nsim, self.lmax+1, self.lmax+1), order='F')
else:
almEs = almEs_in
almBs = almBs_in
if(self.niter == 0):
fastSHT.qu2eb(Q, U, almEs, almBs)
elif(method == 'old'):
fastSHT.qu2eb_iter_old(Q, U, almEs, almBs, self.niter, 0)
if(almEs_in is None or almBs_in is None):
return (almEs, almBs)
else:
return 0
def qu2e(self, Q, U, almEs_in=None):
if(almEs_in is None):
try:
almEs = numba.cuda.pinned_array((self.nsim, self.lmax+1, self.lmax+1), dtype=np.double, strides=None, order='F')
except:
almEs = np.ones((self.nsim, self.lmax+1, self.lmax+1), order='F')
else:
almEs = almEs_in
fastSHT.qu2e(Q, U, almEs)
if(almEs_in is None):
return almEs
else:
return 0
def qu2b(self, Q, U, almBs_in=None):
if(almBs_in is None):
try:
almBs = numba.cuda.pinned_array((self.nsim, self.lmax+1, self.lmax+1), dtype=np.double, strides=None, order='F')
except:
almBs = np.ones((self.nsim, self.lmax+1, self.lmax+1), order='F')
else:
almBs = almBs_in
fastSHT.qu2b(Q, U, almBs)
if(almBs_in is None):
return almBs
else:
return 0
def eb2qu(self, almEs, almBs):
try:
Q = numba.cuda.pinned_array((self.npix, self.nsim), dtype=np.double, strides=None, order='F')
U = numba.cuda.pinned_array((self.npix, self.nsim), dtype=np.double, strides=None, order='F')
except:
Q = np.ones((self.npix, self.nsim), order='F')
U = np.ones((self.npix, self.nsim), order='F')
fastSHT.eb2qu(almEs, almBs, Q, U)
return (Q, U)
def fix_eb(self, Q, U, mask):
vid = (np.arange(len(mask))[mask == 1])
nv = len(vid)
Bmap = np.empty((self.npix, self.nsim), order='F')
alms = np.empty((self.nsim, self.lmax+1, self.lmax+1), order='F')
B_template = np.empty((self.npix, self.nsim), order='F')
flag = 0
fastSHT.fix_eb(Q, U, mask, Bmap, alms, B_template, vid, self.niter, nv, flag)
return Bmap
def convert_alm_healpy(self, alms):
alms_hp = np.empty((2, int((self.lmax+1) * (self.lmax+2) / 2) , self.nsim), order='F')
fastSHT.convert_alm_healpy1(alms, alms_hp, self.nsim, self.lmax)
alms_hp = alms_hp[0,:,:] + 1j * alms_hp[1,:,:]
return alms_hp
# get alms from GPU (only use it when gpu is enabled)
def get_alms(self, alms_in=None):
if(alms_in is None):
try:
alms = numba.cuda.pinned_array((self.nsim, self.lmax+1, self.lmax+1), dtype=np.double, strides=None, order='F')
except:
alms = np.ones((self.nsim, self.lmax+1, self.lmax+1), dtype=np.double, order='F')
else:
alms = alms_in
fastSHT.get_alm(alms)
if(alms_in is None):
return alms
else:
return 0